V21.4.1
Version Release Notes
New Complete HBase to TDSQL Boundless Data Migration Solution
This update provides complete "full + incremental" data migration capabilities:
1. Full migration: Supports complete migration of all historical data from the HBase source database to the TDSQL Boundless target database in a single operation, establishing an accurate data baseline for the migration.
2. Incremental migration: With the source database continuously running and business writes occurring, it monitors HBase WAL logs in real time (supporting Put and Delete operations), and synchronizes new and modified data to the target database via the Kafka message queue, achieving smooth, zero-downtime migration. For details, see Migrate from Self-Deployed HBase to TDSQL Boundless. V21.4.0
Version Release Notes
Syntax and Features
Enhanced HBase Ecosystem Compatibility
Primarily supports compatibility with relevant read/write APIs in HBase mode. Mainly includes data writing categories: Mutate (put, delete, append, increment); data reading categories: Get, Scan; data batch operation categories: Multi. For supported data types and APIs in TDSQL Boundless (HBase compatibility mode), see Compatibility with HBase. Provides the HBase Client JAR and HBase Shell tools for connecting to TDSQL Boundless.
Supports basic ecosystem connectivity (Flink Connector, Kafka Connector, Hive Connector).
Unsupported features:
HBase Coprocessor (coprocessor).
Incremental backup and recovery.
Stability
Optimized Routing Information Acquisition Level
By optimizing the routing information acquisition level from Region to RG, network request pressure is reduced, and system performance is improved.
Bug fixes
Fixed the issue where, when the disk is full and user_cf is in cached write mode, delete_region_job could not be scheduled, resulting in the inability to release disk space.